One of the most exciting developments in treating people with epilepsy, since the turn of the century, is a paradigm shift in our understanding of how epileptic seizures are generated. Rather than starting as abrupt, random events, new evidence suggests that seizure generation is probabilistic, with precursors that wax and wane before some synchronizing event triggers clinical seizures. This line of research has given rise to devices to warn of and pre-empt seizures, some now in clinical trials, and promises exciting therapeutic benefits to patients on the horizon. The research also has great potential to dramatically improve our understanding of the mechanisms underlying seizure generation and epileptogenesis, with even more profound clinical implications. Unfortunately, research in this field is significantly hindered by limited access to continuous, high quality, broad-band recordings from humans implanted with intracranial electrodes, and spontaneously seizing animal models of epilepsy. This is because these data are very expensive to acquire, extremely labor intensive, and the process of filtering, removing artifacts, and annotating recordings spanning weeks to months alone is prohibitive for all but the largest and best-funded investigative teams to undertake. This leaves literally hundreds of qualified scientists who would be actively working in this area unable to engage in this research. We propose to construct an international, collaborative database of broad-band, high quality, annotated intracranial data, from humans and spontaneously seizing animal models of epilepsy, centered at the University of Pennsylvania and Mayo Clinic. Data will be collected from the highest quality facilities worldwide, and made available to all investigators: academic, private and industry, for analysis. The database will be presided over by an international Scientific Advisory Board, and will eventually be a self-sustaining facility, funded by fees charged for data access. This effort will be the centerpiece of The International Collaborative Seizure-Prediction Group, a well-established international collaboration between the top laboratories in the world that study seizure generation, and whose meetings are supported by the National Institutes of Health, American Epilepsy Society, and European EEG Societies. This project will coordinate and collaborate openly with a European database of human intracranial recordings for clinical research. We aim to make this database outlined in this proposal a focal point for collaborative research in epilepsy, both basic science and translational, worldwide. Public Health Relevance: This project will make expensive, difficult-to-acquire, high quality data collected from electrodes implanted in patients during clinical care available to researchers world-wide who work on epilepsy. It will allow them to develop new sensors, devices and treatments for epilepsy. It will also help them understand how seizures and epilepsy begin, so that they can develop new treatments to prevent or cure them.